The equalization and rationalization of educational resource allocation is of great significance to the coordinated development of education. The study takes the educational resources of 13 districts and counties in Y city in 2023 as an example, and proposes to use the BP neural network-based educational resource allocation evaluation system to analyze it. The results show that only three districts and counties have “very good” and “good” levels of educational resource allocation. Accordingly, this paper constructs a multi-objective optimization model to improve the level of educational resource allocation, reduce the differences between counties, and improve the utilization rate of educational resources. The weights corresponding to the eight indicators of the educational resource allocation evaluation index system are solved by the entropy weight method, after which the preset values of the three objective functions and the weights accounted for by the eight indicators are brought into the model and the artificial raindrop algorithm is used to find the optimal solution. After finding the optimal solution of educational resource allocation, the BP neural network-based educational resource allocation evaluation system is used again to evaluate it, and at this time, the educational resource allocation of a total of 12 districts and counties belongs to the “very good” and “good” grades. The study shows that the optimization method of educational resource allocation designed in this paper can reasonably plan educational resources and realize the coordinated development of education.